Very often not enough patients are entered and/or the follow-up is insufficient to be able to draw valid conclusions in cancer clinical trials. In this article, we discuss the possibility of pooling the data from two or more trials asking the same or similar questions in order to overcome such problems. How comparable the studies should be for combining their data, in terms of design, patient population, follow-up, and end-points, is discussed in the first part of this paper. Whether these general considerations were completely or partially fulfilled in the two prostatic studies of the EORTC and DAPROCA is the subject of the second part of this article. Problems of interpreting apparently contradictory results, like the superiority of zoladex and flutamide over orchidectomy in terms of time to progression with no clear superiority in terms of overall duration of survival, is also discussed.